Before the patch, the null program:
#include "types.h"
#include "user.h"
int main() {
char *n = 0;
printf(1, "%x\n", *n);
exit();
}get this output:
pid 3 null: trap 6 err 0 on cpu 0 eip 0x7 addr 0x0--kill proc
it's not a page fault(illegal opcode, what the hell?). Increase n before using it:
n++;
printf(1, "%x\n", *n);then it prints out 4c, check where it come from:
$ x86_64-elf-objdump -d null.o
null.o: file format elf32-i386
Disassembly of section .text.startup:
00000000 <main>:
0: 8d 4c 24 04 lea 0x4(%esp),%ecx
4: 83 e4 f0 and $0xfffffff0,%esp
7: ff 71 fc pushl -0x4(%ecx)
a: 55 push %ebp
b: 89 e5 mov %esp,%ebp
d: 51 push %ecx
e: 83 ec 08 sub $0x8,%esp
11: 0f be 05 01 00 00 00 movsbl 0x1,%eax
18: 50 push %eax
19: 68 00 00 00 00 push $0x0
1e: 6a 01 push $0x1
20: e8 fc ff ff ff call 21 <main+0x21>
25: e8 fc ff ff ff call 26 <main+0x26>
xv6 loads the program at address 0. After applying the patch, virtual address starts at PGSIZE(4096). null(0) is not mapped. Now the error should be:
pid 3 null: trap 14 err 4 on cpu 0 eip 0x1000 addr 0x0--kill proc
// traps.h:
#define T_DIVIDE 0 // divide error
#define T_OFLOW 4 // overflow
#define T_ILLOP 6 // illegal opcode
#define T_DEVICE 7 // device not available
#define T_PGFLT 14 // page fault